Cancer in epithelial tissues develops over a rather long period of time after the putative initiating event. During this process, cells accumulate stable heritable alterations, and intermediate "premalignant" states are sometimes recognizable histologically. In order to achieve earlier diagnosis and more accurate prognosis, and to provide a basis for designing cytotoxic agents specific for cancer cells, it is necessary to identify the features that distinguish normal from malignant cells, and premalignant from bening dysplastic cells. Cell culture has always seemed to be the most promising avenue for such investigation, but it has been difficult to relate the results of experimental culture models to the varied porperties of human carcinomas. The rationale behind our reseach is to use cell culture, the nude mouse, and immunochemical and electrophoretic methods to study malignant transformation in the context of the cell's differentiated state. We have studied in detail six human squamous cell carcinomas (SCCs) of the epidermis and of the oral epithelium. In culture, the SCCs retain varying degrees of dependence on feeder fibroblasts for growth, half of the SCCs synthesize a novel small (38K) keratin, and all of them display a maredly reduced rate of commitment to terminal differentiation as compared with normal keratinocytes. The main objective of this renewal proposal are (1) to classify the stages of cervical neoplasia (normal, dysplasia, carcinoma-in-site, and invasive SCC) by their characteristics of growth and differentiation in culture and in nude mice; (2) to determine whether malignant progression of dysplastic lesions of the cervix is preceded by the presence in such lesions of rare cells that either exhibit and SCC-like phenotype, or of cells that readily transform to such phenotype, in culture; (3) to extend our current studies of oral and pharyngeal SCCs in order to determine the range and types of aberrant differentiated function that can occur during malignant transformation of keratinocytes, and (4) to mutagenize cultured keratinocytes, select variants defective in terminal differentiation, and examine the behavior in nude mice.